-
摘要:
飞行品质监控(FOQA)中对超限事件进行监测是识别飞行潜在风险的重要手段之一，但该方法对各超限事件之间的关联性研究不足，无法量化存在相互影响关系的超限风险。为此，提出一种基于快速存取记录器(QAR)数据与关联规则挖掘的着陆阶段超限风险评估方法。对
9799 个航段样本在进近着陆阶段发生各超限事件的频次进行统计，结合专家意见筛选出152～15 m下降率大、着陆坡度大、着陆速度大、着陆俯仰角小、15 m至接地距离远、着陆垂直载荷大6个着陆阶段风险评估指标；基于超限项目间关联性计算风险指标发生的可能性权重，依据风险指标的参数分布计算航班指标偏移量以评价超限风险的严重性，构建基于云模型的着陆超限风险评估模型。提取QAR数据应用该模型进行验证，结果显示风险评估结果具有区分度。所建模型为着陆阶段的超限风险评估提供了一种有效的方法，有利于飞行安全监管和飞行运行优化。-
关键词:
- QAR数据 /
- 关联规则 /
- 着陆超限 /
- PageRank算法 /
- 风险评估
Abstract:The monitoring and analysis of exceedance events within the flight operation quality assurance (FOQA) program is one of the crucial ways to identify potential risks in flights. However, existing methods lack comprehensive consideration for the interrelationships between various exceedance events, thereby failing to systemically quantify the risks associated with several interdependent events. This study introduces a risk assessment method for landing exceedance events by integrating association rule mining and quick access recorder (QAR) data. Firstly, based on the frequency of exceedance events and the findings of an expert survey conducted during the approach and landing phases of 9 799 flight segments, six exceedance events were found to be indicators of risk assessment: high normal acceleration at landing, low pitch at landing, long landing, excessive bank angle below 50 m, high speed below 50 m, and high descent rate (152～15 m). Secondly, we calculated the possibility and severity of the six indicators based on their correlations and deviations, respectively. Subsequently, a risk assessment model for landing was constructed using the cloud model. Lastly, the real QAR data were used to verify the model. The results can differentiate risk levels associated with specific exceedance events. In order to improve flight safety supervision and operational efficiency, this model offers a useful method for evaluating the exceedance risk during the landing phase.
-
Key words:
- QAR data /
- association rule /
- landing exceedance /
- PageRank algorithm /
- risk assessment
-
图 1 基于关联规则的着陆超限风险评估建模过程
Figure 1. Modelling process for landing exceedance risk assessment based on association rule
表 1 进近着陆阶段超限事件统计
Table 1. Statistics of exceedance incidents in landing phase
超限事件 轻度超限阈值 轻度超限次数 严重超限阈值 严重超限次数 合计次数 $ {A}_{1} $: GPWS警告 0 探测到 17 17 $ {A}_{2} $: 610～305 m( 2000 ～1000 ft)下降率大>7.620 m/s 138 >9.144 m/s 26 164 $ {A}_{3} $: 305～152 m( 1000 ～500 ft)下降率大>6.604 m/s 18 >7.620 m/s 3 21 $ {E}_{1} $: 152～15 m(500～50 ft)下降率大 >5.588 m/s 414 >6.604 m/s 8 422 $ {A}_{4} $: 152～61 m(500～200 ft)进近坡度大 >15° 1 >20° 0 1 $ {A}_{5} $: 61～15 m(200～50 ft)进近坡度大 >8° 16 >10° 2 18 $ {E}_{2} $: 着陆坡度大 >4° 696 >6° 23 719 $ {A}_{6} $: 低空大速度 >118.22 m/s 89 >128.5 m/s 22 111 $ {A}_{7} $: 152～15 m(500～50 ft)进近速度大 >(Vref+12.85) m/s 31 >(Vref+15.42) m/s 5 36 $ {E}_{3} $: 着陆速度大 >(Vref+7.71) m/s 167 >(Vref+10.28) m/s 24 191 $ {A}_{8} $: 放起落架晚 <457.5 m 5 <6.604 m/s 0 5 $ {A}_{9} $: 着陆襟翼到位晚 <366 m 12 <5.08 m/s 2 14 $ {A}_{10} $: 着陆俯仰角大 >7.4° 9 >7.3° 0 9 $ {E}_{4} $: 着陆俯仰角小 <1° 132 <0.5° 32 164 $ {A}_{11} $: 接地点近 <224 m 9 <2.52 m/s 0 9 $ {E}_{5} $: 15 m(50 ft)至接地距离远 >762.5 m 5029 >15.24 m/s 403 5432 $ {E}_{6} $: 着陆垂直载荷大 >1.68g 213 >1.89g 13 226 注：g为重力加速度；Vref为着陆参考速度。 
下载: 导出CSV
表 2 云模型特征参数计算
Table 2. Calculation of cloud model parameter characteristics
特征参数 等级区间 [0, x1] [x1, x2] [x2, x3] [x3, x4] [x4, 3] $ {E}_{{\mathrm{x}}} $ 0 $\dfrac{{{x_1} + {x_2}}}{2}$ $\dfrac{{{x_2} + {x_3}}}{2}$ $\dfrac{{{x_3} + {x_4}}}{2}$ 3 $ {E}_{{\mathrm{n}}} $ $\dfrac{{{x_1} - 0}}{{2.355}}$ $\dfrac{{{x_2} - {x_1}}}{{2.355}}$ $\dfrac{{{x_3} - {x_2}}}{{2.355}}$ $\dfrac{{{x_4} - {x_3}}}{{2.355}}$ $\dfrac{{3 - {x_4}}}{{2.355}}$ $ {H}_{{\mathrm{e}}} $ 0.001 0.001 0.001 0.001 0.001
下载: 导出CSV
表 3 超限监控项目间关联规则
Table 3. Association rules among exceedance events
强关联规则 支持度 置信度 提升度 $ {A}_{7} $→$ {E}_{3} $ 0.014 0.44 3.49 $ {A}_{5} $→$ {E}_{1} $ 0.010 0.56 3.05 $ {A}_{1} $→$ {E}_{1} $ 0.010 0.50 2.74 $ {A}_{5} $→$ {E}_{2} $ 0.014 0.78 2.63 $ {E}_{3} $→$ {E}_{4} $ 0.031 0.25 1.85 $ {E}_{4} $→$ {E}_{3} $ 0.031 0.23 1.85 $ {A}_{2} $→$ {E}_{1} $ 0.012 0.24 1.32 $ {E}_{4} $→$ {E}_{5} $ 0.078 0.58 1.01
下载: 导出CSV
表 4 风险指标可能性权重
Table 4. Weight of risk indicator probability
指标 超限事件名称 可能性权值 E1 152～15 m（500～50 ft）下降率大 0.1310 E2 着陆坡度大 0.0480 E3 着陆速度大 0.1143 E4 着陆俯仰角小 0.1310 E5 15 m（50 ft）至接地距离远 0.3294 E6 着陆垂直载荷大 0.2463
下载: 导出CSV
表 5 风险指标严重度计算
Table 5. Calculation of risk indicator severity
风险指标 样本参数名称 期望 标准差 严重度计算公式 $ {E}_{1} $ 152～15 m（500～50 ft）下降率最大值 896.58 72.51 ${S_{\mathrm{E}}}(1) = ({{{\alpha _1} - 896.58}})/{{72.51}}$ $ {E}_{3} $ 着陆阶段速度最大值 149.70 5.11 ${S_{\mathrm{E}}}(3) = ({{{\alpha _3} - 149.70}})/{{5.11}}$ $ {E}_{4} $ 着陆最小俯仰角 2.70 0.76 ${S_{\mathrm{E}}}(4) = ({{2.70 - {\alpha _4}}})/{{0.76}}$ $ {E}_{5} $ 15 m（50 ft）至接地距离 1371.45 211.62 ${S_{\mathrm{E}}}(5) = ({{{\alpha _5} - 1\;371.45}})/{{211.62}}$ $ {E}_{6} $ 着陆垂直载荷最大值 1.3307 0.07 ${S_{\mathrm{E}}}(6) = ({{{\alpha _6} - 1.330\;7}})/{{0.07}}$
下载: 导出CSV
表 6 样本风险等级评估结果
Table 6. Results of samples risk level assessment
样本序号 风险等级隶属度wi 风险等级 无 轻微 低 中 高 1 0.8396 0.1028 0.0221 0.0260 1.868$ \times $10-12 无风险 2 0.4997 0.5038 0.0400 1.382×10−5 1.490×10-42 轻微风险 3 0.6227 0.1325 0.2972 0.0079 1.226×10−21 无风险 4 0.7004 0.0956 0.2384 0.0082 9.208×10−23 无风险 5 0.4441 0.2136 0.3796 0.0142 5.881×10−20 无风险 6 0.2933 0.3726 0.3908 0.0246 7.196×10−20 低风险 7 0.8587 0.1539 0.0397 6.527×10−5 1.307×0−35 无风险 8 0.6706 0.0597 0.3288 0.0237 5.462×10−20 无风险 9 0.3805 0.4420 0.1412 0.0369 1.311×10−14 轻微风险 10 0.7735 0.1346 0.0001 7.926×10−10 2.425×10−66 无风险
下载: 导出CSV
-
[1] AIRBUS. A statistical analysis of commercial aviation accidents 1958-2022: X00D17008863[R]. Blagnac Cedex: AIRBUS, 2023: 27. [2] Boeing Commercial Airplanes. Statistical summary of commercial jet airplane accidents worldwide operations 1959-2021: 23072954[R]. Seattle: Boeing Commercial Airplanes, 2022: 10. [3] 中国民用航空局. 飞行品质监控(FOQA)实施与管理: AC-121/135-FS-2012-45R1[R]. 北京: 中国民用航空局飞行标准司, 2015: 21-25.Civil Aviation Administration of China. Flight quality assurance (FOQA) implementation and management: AC-121/135-FS-2012-45R1[R]. Beijing: Flight Standards Department of Civil Aviation Administration of China, 2015: 21-25(in Chinese). [4] WANG L, GAO S, HONG R Y, et al. Effects of age and flight exposure on flight safety performance: evidence from a large cross-sectional pilot sample[J]. Safety Science, 2023, 165: 106199. doi: 10.1016/j.ssci.2023.106199 [5] AYRA E S, RÍOS INSUA D, CANO J. Bayesian network for managing runway overruns in aviation safety[J]. Journal of Aerospace Information Systems, 2019, 16(12): 546-558. doi: 10.2514/1.I010726 [6] 孙瑞山, 陈雄, 梁妍. 基于高斯混合模型的飞机进近着陆阶段运行异常检测[J]. 安全与环境学报, 2022, 22(3): 1371-1376.SUN R S, CHEN X, LIANG Y. Abnormal operation detection for approach and landing phase based on Gaussian mixture model[J]. Journal of Safety and Environment, 2022, 22(3): 1371-1376(in Chinese). [7] 祁明亮, 邵雪焱, 池宏. QAR超限事件飞行操作风险诊断方法[J]. 北京航空航天大学学报, 2011, 37(10): 1207-1210.QI M L, SHAO X Y, CHI H. Flight operations risk diagnosis method on quick-access-record exceedance[J]. Journal of Beijing University of Aeronautics and Astronautics, 2011, 37(10): 1207-1210(in Chinese). [8] HADJIMICHAEL M, OSBORNE D M, ROSS D, et al. The flight operations risk assessment system[J]. SAE International, 1999: 150-156. [9] 汪磊, 孙瑞山, 吴昌旭, 等. 基于飞行QAR数据的重着陆风险定量评价模型[J]. 中国安全科学学报, 2014, 24(2): 88-92.WANG L, SUN R S, WU C X, et al. A flight QAR data based model for hard landing risk quantitative evaluation[J]. China Safety Science Journal, 2014, 24(2): 88-92(in Chinese). [10] 陈农田, 满永政, 李俊辉. 基于QAR数据的民机高高原进近着陆风险评估方法[J]. 北京航空航天大学学报, 2024, 50(1): 77-85.CHEN N T, MAN Y Z, LI J H. Risk assessment method for civil aircraft approach and landing at high plateau based on QAR data [J]. Journal of Beijing University of Aeronautics and Astronautics, 2024, 50(1): 77-85(in Chinese). [11] 汪磊, 郭世广, 任勇. 基于飞行数据正态云的着陆操作风险评价方法[J]. 安全与环境学报, 2019, 19(5): 1555-1561.WANG L, GUO S G, REN Y. Landing operation risk evaluation based on the normal cloud of the flight data[J]. Journal of Safety and Environment, 2019, 19(5): 1555-1561(in Chinese). [12] 汪磊, 高杉, 张静怡, 等. 民航飞行员超限行为评价方法研究[J]. 安全与环境学报, 2021, 21(2): 695-700.WANG L, GAO S, ZHANG J Y, et al. Of the airline transport pilots based on the QAR data Evaluation of the exceedance behaviors[J]. Journal of Safety and Environment, 2021, 21(2): 695-700(in Chinese). [13] WANG L, REN Y, WU C X. Effects of flare operation on landing safety: a study based on ANOVA of real flight data[J]. Safety Science, 2018, 102: 14-25. doi: 10.1016/j.ssci.2017.09.027 [14] 谢嘉仪. 基于QAR飞行大数据的不稳定进近风险分析与预警技术研究[D]. 武汉: 武汉大学, 2021: 14-17.XIE J Y. Research on unstable approach risk analysis and early warning technology based on QAR flight big data[D]. Wuhan: Wuhan University, 2021: 14-17(in Chinese). [15] MONTELLA A. Identifying crash contributory factors at urban round abouts and using association rules to explore their relationships to different crash types[J]. Accident Analysis & Prevention, 2011, 43(4): 1451-1463. [16] LI L, GUO H M, CHENG L H, et al. Research on causes of coal mine gas explosion accidents based on association rule[J]. Journal of Loss Prevention in the Process Industries, 2022, 80: 104879. doi: 10.1016/j.jlp.2022.104879 [17] LAN H, MA X X, MA L H, et al. Pattern investigation of total loss maritime accidents based on association rule mining[J]. Reliability Engineering & System Safety, 2023, 229: 108893. [18] 时统宇, 王岩韬, 李鸿坤. 基于先验频繁模式算法的航班延误关联规则挖掘[J]. 飞行力学, 2023, 41(2): 87-94.SHI T Y, WANG Y T, LI H K. Flight delay association rules mining based on Apriori algorithm[J]. Flight Dynamics, 2023, 41(2): 87-94(in Chinese). [19] 袁乐平, 张文东, 赵力梵, 等. 基于TEM模型的民航不安全事件关联规则挖掘与分析[J]. 中国安全生产科学技术, 2022, 18(10): 31-36.YUAN L P, ZHANG W D, ZHAO L F, et al. Mining and analysis of association rules in civil aviation unsafe incidents based on TEM model[J]. Journal of Safety Science and Technology, 2022, 18(10): 31-36(in Chinese). [20] 王明涛. 证券投资风险本质属性探讨及其计量模型研究[J]. 河南金融管理干部学院学报, 2003, 21(4): 61-63.WANG M T. An inquiry into essential attributes of securities investment risks and its quantitative pattern research[J]. Journal of Henan College of Financial Management Cadres, 2003, 21(4): 61-63(in Chinese). [21] AGRAWAL R, IMIELIŃSKI T, SWAMI A. Mining association rules between sets of items in large databases[J]. ACM SIGMOD Record, 1993, 22(2): 207-216. doi: 10.1145/170036.170072 [22] 毕建欣, 张岐山. 关联规则挖掘算法综述[J]. 中国工程科学, 2005, 7(4): 88-94. doi: 10.3969/j.issn.1009-1742.2005.04.016BI J X, ZHANG Q S. Survey of the algorithms on association rule mining[J]. Strategic Study of CAE, 2005, 7(4): 88-94(in Chinese). doi: 10.3969/j.issn.1009-1742.2005.04.016 [23] 蔡伟杰, 张晓辉, 朱建秋, 等. 关联规则挖掘综述[J]. 计算机工程, 2001, 27(5): 31-33. doi: 10.3969/j.issn.1000-3428.2001.05.014CAI W J, ZHANG X H, ZHU J Q, et al. Survey of association rule generation[J]. Computer Engineering, 2001, 27(5): 31-33(in Chinese). doi: 10.3969/j.issn.1000-3428.2001.05.014 [24] BRIN S, PAGE L. The anatomy of a large-scale hypertextual web search engine[J]. Computer networks and ISDN systems, 1998, 30(1-7): 107-117. [25] 李德毅, 杜鹢. 不确定性人工智能[M]. 2版. 北京: 国防工业出版社, 2014: 42-43.LI D Y, DU Y. Artificial intelligence with uncertainty[M]. 2nd ed. Beijing: National Defense Industry Press, 2014: 42-43(in Chinese). [26] 叶琼, 李绍稳, 张友华, 等. 云模型及应用综述[J]. 计算机工程与设计, 2011, 32(12): 4198-4201.YE Q, LI S W, ZHANG Y H, et al. Cloud model and application overview[J]. Computer Engineering and Design, 2011, 32(12): 4198-4201(in Chinese). -
下载:
点击查看大图
计量
- 文章访问数: 21
- HTML全文浏览量: 2
- PDF下载量: 4
- 被引次数: 0
